Markdownee Apify Actor

Use the Actor for hosted crawls, schedules, and Apify storage. Start with one page, inspect its output, then expand the scope.

Start a run

Paste this into the Actor's Input tab and select Start, or save it as input.json for the commands below:

{
  "startUrls": [{ "url": "https://en.wikipedia.org/wiki/Web_scraping" }],
  "crawlerType": "cheerio",
  "maxRequestsPerCrawl": 1,
  "save": ["markdown-kvs"]
}

Inspect the successful run's Dataset record and its Markdown file in the Key-value store. Add a link selector, URL filters, and limits to follow links; globs alone discover nothing. The Input tab is the generated reference for all fields.

For command-line or API use, set APIFY_TOKEN to your Apify API token in the execution environment. The CLI route requires Node.js/npm; the API route requires curl.

npx --package=apify-cli apify call markdownee/crawler \
  --input-file input.json --output-dataset

This waits for the run and prints its default dataset. To start a run through the API:

curl --fail --request POST \
  "https://api.apify.com/v2/acts/markdownee~crawler/runs" \
  --header "Authorization: Bearer $APIFY_TOKEN" \
  --header "Content-Type: application/json" \
  --data-binary @input.json

The API returns run metadata immediately; use its run and store IDs to inspect status and retrieve results. See the API tab. Keep tokens out of shared files and URLs.

Input reference

A blank default means the schema has no fixed value. Settings apply only to their supported crawler or storage surface.

Crawl settings

Crawl filtering

Content extraction

Output destinations

Browser settings

Proxy

Cookies & headers

Read outcomes and content

Dataset records use success, failed, or, when enabled, skipped status. Failures include errors and retry counts; skips include skipReason. Each record holds the request URL, metadata, and available crawl context such as loaded URL, timestamp, HTTP status, and depth.

A *-dataset route includes content inline; *-kvs stores a file and references its key and available public URL. Content nodes carry MD5 hashes and UTF-8 byte counts. Select both destinations to keep both representations; large HTML bodies are better kept in KVS. The Output tab defines the record schema.

Format selectors use minified-html; records use minifiedHtml. The original node describes captured input, and an original-* route stores its bytes. That capture may already reflect browser rendering or consent handling.

When discovery accepts origin-published Markdown, markdownSource records the URL, mechanism, and whether the saved Markdown is verbatim. Its original HTML is then derived from that representation. Discovery depends on crawler capabilities and request budgets; rejected representations fall back to HTML. A Markdown-only response has no page HTML from which to follow links.

See CLI Help and npm library Help for local execution, or Python Help for the native library.

Updated: September 25, 2026